Effective electronic health records

The staff of the Hôpital du Valais works with a large number of different software on a daily basis. For the nurses and physicians, the main touchpoint is the electronic medical record (EMR). The EMR-software was purchased from an external vendor but offered the possibility to create custom forms with a form-builder. In addition to that, it was possible to plug web applications into the software. In March 2018 I joined the IT department as their first designer to help design these applications.

My role

As the only designer in a team of developers, I was responsible for field research, usability testing, UX and UI design of the web application and writing specifications for the development.


A new web application that replaces numerous static forms and makes data accessible across different wards and clinics.

Understanding pain points of different user groups

So far there was no habit of field research at the hospital. Requirements for new software were usually defined in hour-long meetings between developers and user-representatives. This time luckily, we got permission to observe people at their workplace in addition to conducting interviews. This helped us uncover the following problems:

  • Software applications are designed for silos – Crucial patient information is not shared between different departments or professional groups, resulting in a lot of re-written documentation.
  • Unstructured data makes it impossible to leverage information – Doctors have a habit of dictating or writing medical information in the form of long paragraphs, which makes it impossible to reuse specific parts of the data.
  • Patients are excluded from there medical records – Especially elderly patients are struggling with handling medical information as well as remembering doctors’ appointments.
A user journey map showing pain points a patient might experience when visiting the hospital.
A user journey map visualising inefficient steps in the workflow of a secretary.

Defining the scope

Until this point, any staff member of the Hôpital du Valais could request the development of a new form. This process resulted in an unmanageable backlog and hundreds of forms each designed for only a small group of users. I mapped out all existing forms in order to discover which had overlapping functionalities that could be covered by one single web application.

Technological limitations and the lack of a product design process led to hundreds of forms.

Getting more out of data

The current forms oblige doctors and nurses to input the same data twice or even three times. At the same time the data is hardly ever re-used in different contexts, statistics or automated summaries.

We decided to work with FHIR because it would allow us to define relationships between different data points in the patients’ medical records. For example, if a transfusion reaction is documented on the transfusion form, it could be displayed automatically as a “Complication” on the problem list.

FHIR (Fast Healthcare Interoperability Resources) is a data standard for electronic health records.

Selecting a UI framework

Together with the developers, I decided to work with Google Material Design and Angular Material as a framework. Some of the reasons that supported this decision were that MD had been designed with desktop and touch use in mind, it’s very widely adopted and it would speed up the development process.

Defining standards

A quick-win which would potentially have a big impact on the productivity of the medical staff, was to standardize the way in which recuring information was displayed.

I created a layout with header, including page related tasks such as ‘print’ and ‘help’, and sidebar, hosting crucial patient information like ‘allergies’ and ‘reanimation status’.

The header contains recurring functionalities, like a help page and a modification history.
The sidebar contains the most important information about a patient and can be extended or closed depending on screen size.

The solution

This research and design project led to a pipeline of web applications, that will be developed and implemented gradually. The main improvements we achieved were:

  • Fewer applications for more users – Thanks to filtering, search functions and better UI design there is no more need to have separate forms for each user group. Having different user groups use the same application reduces the need to double documentation and improves knowledge sharing among disciplines.
  • Meaningful data connections – Reusing data in different contexts increases the value of a single data input and saves time.
  • Interaction design standards to minimize cognitive burden – Allowing users to find information easily and always in the same place reduces frustration and onboarding time when changing jobs within the hospital.
  • Easy access to help – Allowing users to find information on how to use the applications cuts the need for IT support, which frees resources for further product development.

Hospital stay

  • Gives a high-level overview over the patients current visit to the hospital.
  • Displays the patients journey through the hospital and allows to retrieve relevant information.
  • Has a modular design that allows to adapt the UI to the specific needs of different departments.

Problem list

  • A list of the patients conditions that is shared across departments and hospital visits.
  • Drag and drop allows to quickly change the category of a problem.


  • A chronological feed of notes about a patient.
  • Notes can be searched or filtered by user group.


  • A searchable table with all interventions a patient had during different stays at the hospital.
  • Interventions can be connected to conditions in order to avoid documenting the same information twice.
  • An underlying catalogue automates steps of the billing process.


There is a big difference between what users want and what they need. Developing every request from a user results in feature creep and thus bad products. Talking to people is an important part of human centered design, but equaly so a phase of analysis and interpreting user statements.